MARINE ECOLOGY PROGRESS SERIES Vol. 183: 169-178.1999 1 Published July 6 Mar Ecol Prog Ser l

Distributions of total and active bacteria in biofilms lining tubes of the onuphid polychaete Diopatra cuprea

Tina M. Phillips, Charles R. Lovell*

Department of Biological Sciences, University of South Carolina, Columbia, South Carolina 29208, USA

ABSTRACT: Infaunal burrows and tubes support elevated levels of microbial biomass and activities relative to the surrounding bulk sediments. The impacts of these structures on sediment biogeochem- istry have been intensively studled, but very little is known about the microbiota or their spatial orga- nization within these microenvironments. Distributions of bacterial cells and microcolonies and of potentially active bacterial cells in tubes of the onuphid polychaete Diopatra cuprea were determined using confocal scanning laser microscopy. Distributions were examined among different tubes, at dif- ferent locations within a tube, and at different depths in the biofilms lining these tubes. The average quantity of cells across all tubes examined was 5.61 X 10' cm-3 and there was no trend in the distribu- tion of cells along the length of the tube or with increasing depth in the biofilm. Cell size across all tubes collected in February 1997 averaged 0.085 pm3 but varied seasonally. Mean cell size increased with increasing depth below the sediment surface of the tube site sampled, and decreased with increasing depth in the biofilm. Microcolonies were abundant and not confined to specific depths within the biofh or locations within the tube. Potentially active cells were observed at all depths within the biofilm and at least 46% (mean = 68%) of cells at any given site were potentially active. We hypothe- size that the observed trends in cell distribution are responsive to tube irngation by the host polychaete

KEY WORDS: Microbial communities . Infaunal burrows . Spatial distributions . Confocal scanmng laser microscopy

INTRODUCTION 1980, Grant 1983).Sediment mixing can greatly stimu- late some microbial activities by increasing the quanti- Geochemical cycling of the major biogenic elements ties of potential electron donors and acceptors below in nearshore marine sediments is primarily mediated the sediment-water interface (Findlay & White 1983, by microorganisms. However, sediments are far from 1984, Krantzberg 1985, Findlay et al. 1990). Mixing homogeneous and some specific microenvironments also disrupts chemocline microbial communities and are particularly conducive to microbial activities. may significantly alter the interactions among various Chemocline environments, such as those found at the microbial functional groups and affect their potential sediment-water interface, are well known to be sites of for consortial metabolism (Paerl & Pinckney 1996). enhanced microbial activities (Novitsky 1983, Novitsky Macrofaunal burrows and tubes also provide signifi- & Karl 1986). The classic model of these interface envi- cant chemocline niches for microbial growth. These ronments places aerobic metabolism at the sediment structures are dug into anoxic subsurface sediments, surface, followed by various anaerobic processes in irrigated with oxic seawater, and provide a large sur- deeper sediment layers (e.g. Fenchel 1969, Revsbech & face area for diffusive exchange (Aller & Yingst 1978, Jorgensen 1986, Aller 1988). However, surface sedi- Aller 1980, Boudreau & Marinelli 1994). Irrigation of ments are subject to high levels of disturbance due to the burrow by the macrofaunal host organism results physical and biogenic reworking (Yingst & Rhoads in the formation of radial chemoclines extending out from the burrow wall into the surrounding sediments 'Addressee for correspondence. E-mall: [email protected] (Aller & Yingst 1978, Aller 1983, 1988, Boudreau &

O Inter-Research 1999 Resale of full article not permitted 170 Mar Ecol Prog Ser 183: 169-178, 1999

burrow lining biofilms differs from that of adjacent bulk sediments (Dobbs & Guckert 1988, Steward et al. 1996) and that burrow and tube microbiota play impor- tant roles in the mineralization of organic compounds and the geochemical cycling of biogenic elements (e.g. Kristensen et al. 1985, 1991, Binnerup et al. 1992). The rates of geochemical processes within the burrow biofilm rely mainly on the constituents of the microbial community and their levels of activity. The latter are obviously hlghly dependent on the properties of the burrow (stability and permeability) and on the irriga- tion and exudation activities of the host animal. How- ever, the distributions and potential for consortial interactions of these organisms are almost completely unknown and can best be studied using microscopic analysis. Recent developments in microscopy, specifically confocal scanning laser microscopy, allow examination of bacterial biofilms in a non-invasive manner (Coster- ton et al. 1995). Studies to date have primarily exam- ined biofilms grown on impermeable surfaces and little is known about natural biofilms growing on permeable surfaces and the rnicrobiota inhabiting them. In this study, we examined distributions of total and active bacterial cells in biofilms lining the stable tube struc- ture constructed by a ubiquitous infaunal polychaete, Diopatra cuprea. Several key features of the tube lin- ing biofilm microbiota were revealed. Fig. 1. Diopatra cuprea tube structure and the sampling pro- cedure used to determine distributions of bacteria in D. cuprea tube lining biofilms. For each tube, two 8 cm segments CHARACTERISTICS OF THEDIOPATRA CUPREA of the unreinforced portion (Segments A and B) were col- lected. In each of the segments, 5 sites (1 to 5) were examined. TUBE At each site, confocal images were collected at 1 pm depth intervals from the surface of the biofilm to the tube wall The Diopatra cuprea tube is constructed from a mucous, sulphated polysaccharide, released from secreting glands, which hardens when it comes in con- Marinelli 1994, Marine111 & Boudreau 1996). These tact with seawater (Myers 1972). Each tube displays 2 radial chemoclines provide overlapping gradients of types of construction, reinforced and unreinforced potential electron donors and acceptors and are stable (Fig. 1). The reinforced section rises from the sediment over time spans long enough for development of surface and can also extend down into shallow subsur- complex microbial communi.ties (Aller & Aller 1986, face sediments. The length of this section is dependent Marinelli et al. unpubl.). Macrofaunal burrows vary on the degree of sediment disturbance, and the poly- in their structural properties, ranging from loosely chaete will extend it if the exposed portion becomes packed sediment/mucopolysaccharide aggregates, as buried. The reinforced section of the tube is decorated seen in the burrows of the terebellid polychaete with debris (small pebbles, shell fragments, and plant Amphjtnte ornata, to well-defined polysaccharide detritus) from the sediment surface. Occasionally, tubes and/or proteinaceous tubes, such as those of the will have 2 reinforced openings. The unreinforced sec- chaetopterid polychaete Chaetopterus variopedatus tion occurs below the sediment surface and is con- and the onuphid polychaete Diopatra cuprea (Myers structed as the worm moves vertically down through 1972, Aller & Yingst 1978, Aller 1983).These structures the sediments, creating a mucous sheath. The only for- promote development of microbial biofilms containing eign materials that may be attached to this portion of high levels of microbial biomass (Steward et al. 1996) the tube are sediment particles in direct contact with and supporting intense geochemical cycling activities the mucus when it is secreted. Additional layers of (Kristensen et al. 1985, 1991, Binnerup et al. 1992). mucus are laid within the tube over time, so the thick- Previous studies have shown that the microbiota of ness of the tube is hrectly proportional to its age. Phillips & Lovell: Bacterial distributions in Djopatra cuprea tubes 171

MATERIALS AND METHODS These samples were transported to Columbia, SC, and held in a recirculating sea water tank overnight. Sampling site. Diopatra cuprea tubes were collected Sample preparation and staining procedures. Each from an intertidal sandflat located at Oyster Landing, of the 10 tube segments from February 1997 was in the North Inlet saltmarsh (33" N, 79" W), near secured onto a glass microscope slide with rubber Georgetown, South Carolina, USA. Oyster Landing is a bands approximately 5 cm apart. The 2 cm ends desig- small tidal tributary bordered by Spartina alterniflora nated for handling were then removed. The portion of marsh and oyster beds (Allen et al. 1992, Lovell & the 8 cm segment between the 2 rubber bands was cut Piceno 1994). The sandflat from which samples were longitudinally and unrolled with the inner tube surface collected is completely exposed during low tide and facing up. These samples were stained for enumera- consists of quartz sand (median grain size of 376 pm) tion and measurement of total cells and microcolonies. containing 0.22% silt and clay, and 0.83% organic SYTOX Green was selected as the primary fluores- matter, by dry weight (Lovell & Piceno 1994). cent stain after substantial background fluorescence Sample collection. All Diopatra cuprea tubes col- problems were encountered with acridine orange and lected for this study were built by solitary individuals fluorescein isothiocyanate. DAPI (4,6-diamidino-2- and had no tube branches. The amount of debris and phenylindole) was not tested since the excitation firm attachment of the debris particles to the tube in wavelength needed for this dye could not be obtained the reinforced section made it impossible to mount with our confocal microscope system. Fluorescence samples of intact reinforced segments for microscopic background from SYTOX Green, SYTO-9, and propid- examination or to remove the debris without dis- ium iodide was quite low and permitted detailed ruption of the inner lining biofilm. Therefore, only assessment of cell distributions and appearances. Of a the unreinforced sections of the tubes were analyzed 12.5 pM solution, 100 p1 of SYTOX Green (Molecular in this study. The tube caps (apex of the reinforced Probes, Inc.,Eugene, OR) was added directly onto the section) were cut off at the sediment surface and the inner surface and the sample incubated at room tem- tubes exhumed until approximately 20 to 25 cm of perature in the dark for 5 min. SYTOX Green pene- the unreinforced portions were exposed. The length trates only fixed cells (Roth et al. 1997), complexing of the reinforced portion of each tube was measured with DNA. The stain-DNA complex fluoresces green and the approximate distance of the upper end of the when excited (excitation/emission: 504/523 nm). After unreinforced section from the sediment surface de- incubation, a solution of 1,4-diazobicyclo[2,2,2]oc- termined. From each tube a 20 cm sample (the deep- tane:glycerol (1:3) (Sigma, St. Louis, MO) was added to est 2 cm of the reinforced and the first 18 cm of the reduce photobleaching. A 22 X 40 mm coverslip was unreinforced) was removed and further dissected into taped down on top of the sample, holding it flat. The two 10 cm segments, designated Segments A and B unfixed October 1997 tube segments were mounted as (Fig. 1). These segments were selected to permit described above and stained for the enumeration of examination of the biofilm at close proximity to the active and inactive cells (Lawrence et al. 1997). Meta- tube inlet and at greater distance from it. Segment A bolically active bacteria create a proton gradient consisted of 2 cm of the reinforced and the first 8 cm across their plasma membrane through active trans- of the unreinforced. Segment B was the remaining port of protons out of the cell. This electrochemical 10 cm of the unreinforced tube sample. The 2 cm gradient, called the proton motive force, or membrane reinforced end of Segment A and the first 2 cm of potential, provides a source of energy that can be used Segment B permitted convenient handling of the to drive ATP synthesis. SYTO-9 (excitation/emission: tube samples without disruption of the biofilm in the 488/509 nm; Molecular Probes, Inc.) only penetrates remaining 8 cm segments. cells having membrane potential (Lebaron et al. 1998) Five Diopatra cuprea tubes were collected in Febru- and fluoresces green upon excitation. These cells are ary 1997. The water and sediment temperatures at the considered to be potentially active. Propidium iodide time of collection were 20 and lg°C, respectively. The (excitation/emission: 535/617 nm; Molecular Probes, samples were placed in 0.2 pm filtered 2.5% formalin, Inc.) only penetrates cells lacking membrane potential transported to Columbia, SC, and stored at 4°C in the (Kaneshiro et al. 1993) and fluoresces red upon excita- dark for no more than 1 wk prior to analysis. Five addi- tion. These cells are considered to be inactive. Of a tional D. cuprea tubes were collected in October 1997. 50 SYTO-9, 149 pM propidium iodide staining The water temperature was 26'C and the sediment solution 100 p1 was added to the samples, which were temperature was 25°C at that time. Tubes were han- incubated and mounted as described above. dled as described above, but with the following Confocal scanning laser microscopy. A Bio-Rad modifications: (1) only Segment A for each tube was MRC-1000 Laser Scanning Confocal Microscope (Her- collected and (2) tube samples were not formalin fixed. cules, CA) equipped with a kryptodargon ion laser 172 Mar Ecol Prog Ser 183: 169-178, 1999

was used to produce fluorescence images of the tube vals were indicated by yellow highlighting and were lining biofilms. Confocal microscopy allows optical counted only in the first of the 2 intervals. Bacterial sectioning of the biofilm using an adjustable x-axis cells present in multicellular aggregates of more than stepping motor. At set successive intervals, the tunable 5 cells were not counted in these profiles. Lengths and laser (set to the appropriate excitation wavelength) is widths of 10 arbitrarily selected cells were determined scanned over the sample and excites fluorescent dye for each depth interval by optical magnification of col- complexes in a specific focal plane, allowing detection lected images and measurement using calipers. If less of stained cells at that specific depth (2)in the biofilm. than 10 cells were present in an interval, then each cell Data were collected at 1 pm depth intervals in the was measured. Mean cell size (average biovolume of a biofilm until the tube wall was encountered, providing single cell) for each interval was calculated from the a complete z-profile of images from the top of the length and width measurements determined for single biofilm to the tube wall. The laser does not penetrate cells (Norland 1993). The total biovolume of single below the tube wall inner surface, and no fluorescence cells for each 5 pm interval was calculated by multiply- is observable below that depth (Fig. 1). ing the number of cells in that interval by the mean cell Two z-profiles were taken at each of 5 sites in each of size for that interval. the SYTOX Green stained tube segments. One profile Fifty sites were analyzed for the distribution of all of each pair was collected at lOOOx magnification to types of multicellular aggregates. All multicellular allow determination of single cell distribution, the aggregates, whether tightly packed and representing other at 400x magnification for determination of micro- relatively high densities of cells, or loosely arranged, colony distribution. The first site in each tube segment with relatively few cells, were designated as micro- was approximately 2.5 cm from the top of the segment colonies. For each site, the 1 pm images collected at (Site l), with successive sites (Sites 2 to 5) at 5 mm 400x magnification were compiled in 3 pm intervals. intervals along the length of the segment. The dimen- Some sites did not contain microcolonies. In some sions of the fields of view were dependent on the others, profiles were not completely perpendicular to magnification used (1000~:114 X 76 pm, 400x: 288 X the tube wall, prohibiting accurate measurement of 192 pm). Although data sets for single cell and micro- depths in the biofilm. A total of 46 sites yielded useful colony distributions were collected at approximately data on microcolony distributions and sizes. Micro- the same sites, each is considered to be an indepen- colony cross sectional area in each of the 3 pm inter- dent sample due to microscale differences in biofilm vals was measured using MorphoSys (Meacham & surface texture and the difference in the areas ana- Duncan 1991). Biovolumes of the microcolonies were lyzed using 1000x and 400x magification. estimated by multiplying the microcolony cross sec- Profiles were also collected at 5 sites at lOOOx mag- tional areas in the sequential 3 pm intervals by the nification in each of the activity stained samples. A total depth of biofilm through which the colonies split spectral line (488 and 568 nm) was used to excite extended. No attempt was made to differentiate bio- both SYTO-9 and propidium iodide in order to visual- volume occupied by cells from that containing only ize the potentially active (green) and non-active (red) exopolymeric materials. Due to microscale differences cells simultaneously. Data were not collected on micro- in the sites at which single cell and microcolony data colonles in segments stained to determine potentially were collected, total biovolume of bacteria, which active bacteria due to the bright fluorescence of SYTO- would include all cells regardless of differing modes 9, which tended to overwhelm the fluorescence of pro- of growth, was not calculated. pidium iodide in multicellular aggregates. Depth intervals (5 pm) of each profile at the 5 sites in Data collection. Data collected at 1 pm depth inter- each of the tube segments stained to determine poten- vals in the biofilm to determ~nethe distribution of sm- tially active bacteria were compiled as described gle bacterial cells (February 1997 tubes) were com- above and the number of potentially active cells and piled into a series of 5 pm intervals (i.e. 0-5 pm, non-viable or inactive cells in each of the resulting 6-10 pm, 11 -15 pm, etc.) and the number of cells pre- 5 pm depth intervals were counted. Duplicate counting sent in each interval counted. A total of 50 sites were of cells located at the border of 2 adjacent intervals was examined for single cell distribution within the 5 tubes, avoided by aligning the 2 intervals side by side and but data for 1 site in 1 tube (Tube 1, Segment A, Site 1) comparing cell placement in the field of view. One were lost due to faulty digital storage. In order to avoid arbitrarily selected profile from each activity stained duplicate counting of cells that were at the border of tube segment was used for determination of average 2 adjacent 5 pm intervals, each interval image (except sizes of active and inactive cells (Tube 1, Site 1; Tube 2, 0 to 5 pm) of a sequential pair of intervals was high- Site 1; Tube 3, Site 5; Tube 4, Site 2;Tube 5, Site 3) as lighted in a different color (red or green) and the described above. Lengths and widths of all potentially 2 images were overlapped. Cells present in both inter- actwe and inactive cells were measured and mean cell Phillips & Lovell: Bacterial dlstrlbutions in Djopatra cuprea tubes 173

sizes calculated for each of the 5 pm intervals in these profiles as described above. A Statistical analyses. The SAS System for Windows (version 6.12, SAS Institute, Inc., Cary, NC, USA) was used to perform all statistical analyses. Nested analysis site 5- of variance was used to determine significant differ- SegmentB - ences in cell numbers, cell size, and biovolume among site 1 - tubes, among segments, and among sites for data col- lected from Diopatra cuprea tubes stained with SYTOX Green. Regression analysis was used to evalu- ate relationships between cell numbers, mean cell size, or biovolume, and depth in the biofilm for all sets of 0 2 4 6 8 10 12 14 tube segments. Cell size variance was modeled using the gamma distribution with generalized linear model, Cell numbers (10* l cm3) using generalized estimating equations to accommo- SegmentA date the nested design. All data were tested for nor- site 1 mality and statistical tests were performed using a sig- site 2 nificance level of a = 0.05. site 3 site 4 site 5 RESULTS AND DISCUSSION SegmentB site 1 Bacterial single cell distributions and sizes site 2 site 3 site 4 Confocal microscopy revealed that the tubes of site 5 Diopatra cuprea supported substantial biofilms con- taining abundant bacteria occurring as single cells and in microcolonial formations. The thickness of D. Mean cell size (pm3) cuprea tube biofilm ranged from 30 to 110 pm, with an average thickness of 60 pm. The quantity of indi- SegmentA vidual bacterial cells at a given site ranged from 1.38 site l - X 10' to 1.33 X log cells cm-3 (mean = 5.61 X 10' cells site 2. C cm-3, Fig. 2A). Nested analysis of variance revealed site 3. - that cell numbers differed significantly from site to site4. - site (df 8, 31; F = 3.00; p 5 0.02), but not among tubes site 5. W (df 4,4; F= 3.14; p 2 0.14) or between Segments A and SegmentB . B (df 4,8; F= 1.01; p 2 0.45). There was no consistent site 1 . site 2- pattern of cell distribution within the tube lining ++ site 3- biofilm (Fig. 3). Twenty-five of the 49 sites had high- site 4- I- I est cell numbers within 10 pm of the biofilm surface. site 5- 1 - l Twenty-three had additional subsurface peaks in cell - numbers. Cell numbers decreased significantly (regres- 0 5 10 15 20 25 30 sion analysis, p < 0.05 in all cases) with increasing Biovolume (1o6 pm3 1cm3) depth in the biofilm at 28 of the 49 sites, but the remaining sites showed no significant changes with Fig. 2. (A) Cells per cm3, (B)mean cell size, and (C) biovolume (cell numbers multiplied by mean cell size at each site) in depth. Diopatra cuprea tube lining biofilrns. Mean (* SD) is shown Unlike cell numbers, mean bacterial cell sizes did for each of the 5 sites within each segment across all 5 tubes change with depth below the sediment surface. Cell (n = 5 except for Segment A, Site 1, where n = 4) sizes at different sites in the tube segments ranged from 0.032 to 0.285 pm3 (grand average: 0.085 pm3, Fig. 2B). Nested analysis of variance revealed signifi- among tubes (df 4,4; F= 0.60; p 2 0.68) or sites (df 8,31; cant differences in mean cell size between Segments A F = 1.19; p L 0.3). Based on Diopatra cuprea behavior and B (df 4,8; F = 4.93; p 5 0.03), with cells in Seg- and bacterial stress responses, this trend is quite dif- ment B averaging larger than those in Segment A. No ferent from what would be expected. D. cuprea spends significant differences in mean cell size were found most of its time in the upper portion of the tube neal- 174 Mar Ecol Prog Ser 183: 169-178, 1999

Mean cell size (pm3) Mean cell size (prn3)

0 2 4 6 8 10 12 14 16 8 3 Cell numbers (10 /cm ) 0-5 6-10 11-15 Biovolurne (106 pm31cm3) 16-20 2 1-25 26-30 31-35 36-40 41-45 46-50 51-55

0 2 4 6 8 10 12 14 16 Fig. 3. Distribution of cell numbers (bars), mean cell size Cell numbers (1 08 / cm3) (fSD, n = 1 to 10) (0) and biovolume (cell numbers multiplied I I I I pm )24681012 ,4 ,L by mean cell size per 5 interval) (A) with increasing depth into Diopatra cuprea tube lining biofilm for Sites 1 (A),2 (B). Biovolurne (I o6prn3/~m3) 3 (C),4 (D)and 5 (E)in Segment A of Tu.be 2 the sediment surface and irrigation of the tube is fre- counterparts (Kjelleberg et al. 1987, Chesbro et al. quent but not continuous. At the depths of the tube 1990, Kaprelyants et al. 1993). If cells located deeper in sites examined (10 to 36 cm below the sediment sur- the tube were oxygen and/or nutrient limited they face), the only oxygen available to the rnicrobiota and should have been smaller than those at shallower a substantial portion of the nutrient inputs are pro- depths, but that was not the case. vided by host irrigation. Consequently, sites at greater There was also a clear pattern in the distribution ot depths should experience greater oxygen and/or nutri- bacterial cell size with depth in the burrow hning biofilm ent limitation due to the lesser impact of irrigation. (Fig. 3).The largest mean celI sizes were found in inter- Bacteria in nature that are severely nutrient limited are vals within 10 pm of the biofilm surface at 32 of the typically smaller than their more resource-sufficient 49 sites. Nineteen sites had subsurface peaks in cell size. Phillips & Lovell: Bacterial distributions in Diopatra cuprea tubes 175

Almost all sites (42 out of 49) showed significant 0.21), or among tubes (df 4,4; F = 1.96; p t 0.26). The decreases in mean cell size (regression analysis, p < 0.05 structures of microcolonies varied from very tightly in all cases) and of variance in cell size (E = -0.0005; SE = organized formations to loose aggregations of cells 0.0002; Z = -1.906; p I 0.01) with increasing depth into and exopolymers. The potential of these structures to the biofilm. This trend may have been due to nutrient facilitate interactions among different bacterial spe- or oxygen limitation at the base of the biofilm. cies (Paerl & Pinckney 1996) is clear and their forma- The biovolume of single cells at a given site, a func- tion may be very important to the overall activities of tion of the grand average of cell sizes (i.e. average of the biofilm. mean cell sizes for all intervals at that site) and cell numbers, ranged from 7.26 X 106 to 3.53 X 108 pm3 cm-3 (mean = 5.33 X 107 pm3 cm-3, Fig. 2C). Nested analysis Potentially active versus inactive cells of variance showed that biovolume differed signifi- cantly among sites (df 8,31; F = 2.39; p 10.04) but not Diopatra cuprea tubes were also stained to permit between Segments A and B (df 4,8; F = 2.06; p 2 0.17) determination of the proportions of cells that were or among tubes (df 4,4; F= 0.95; p 2 0.51). The highest intact and potentially active versus those lacking mem- biovolumes in 24 of 49 sites were within 10 pm of the brane potential and thus presumably inactive (Law- biofilm surface while 12 sites had subsurface biovol- rence et al. 1997). The total number of cells per site in ume peaks. There was a significant relationship segments stained to determine bacterial activity between biovolume and depth in the biofilm at 34 of 49 ranged from 2.62 X 10Bto 1.71 X log cells cm-3 (mean = sites, and biovolume decreased with increasing depth 6.11 X 10' cm-3). The proportion of total cells that were in the biofilm in at least half of the sites in each tube potentially active was quite high (Fig. 4), ranging from (Fig. 3). 0.46 to 0.85 (mean = 0.68), and did not significantly dif- fer among sites (df 4,16; F= 1.33; p t 0.29). Potentially active cells were observed at all depths within the Microcolony distribution biofilm (Fig. 5) and the proportion of cells that were potentially active was similar at all depths at 19 of the Microcolonies covered a broad range of sizes and 25 sites. The proportion of active cells increased with were found throughout the length of the Diopatra increasing depth in the biofilm at 3 sites and decreased cuprea tube and at all depths within the biofilm. The with increasing depth at the remaining 3 sites (p < 0.05 number of rnicrocolonies at a given site ranged from in all cases). The abundance of potentially active cells 1 to 31 (mean = 5) and these formations were located was not surprising since previous studies have shown at various depths throughout the biofilm. Microcol- ony biovolume ranged from 16 pm3 to 2.5 X 105 pm3 (mean = 1.28 X 104 pm3). Nested analysis of variance revealed that the microcolony biovolumes did not differ significantly among sites (df 8.28; F = 1.55; p t 0.18), between Segments A and B (df 4,8; F= 1.83; p t

Cell numbers (10* 1cm3) Fig. 5. Total (black bars) and potentially active (open bars) cells per cm3 in Segment A of Diopatra cuprea tube lining Site biofilnls. Mean (+ SD) at each depth in the biofilm, across all Fig. 4. Total (black bars) and potentially achve (open bars) cells sites, is shown. The number of samples varies from 1 to 25 at per cm3 for each site in Segment A of Diopatra cuprea tube different depths in the biofilm due to differences in biofilrn Lining biofllms. Mean (+ SD) is shown for each site for 5 tubes thickness at different sites 176 Mar Ecol Prog Ser 183: 169-178. 1999

nificantly larger than that of inactive cells (range: 0.15 to 0.28 pm3, grand mean = 0.18 pm3) across all 5 tubes (df 4; T = 4.37; p 2 0.01) (Fig. 6). There was no general pattern in the distribution of potentially actlve or inac- tive mean cell sizes in relation to the depth in the biofilm (Fig. 7). The variance of potentially active cell size significantly decreased with increasing depth in the biofilm (E = -0.0017; SE = 0.0006; Z = -2.945; p 5 0.01) while the variance of inactive cell size increased with depth (E = 0.0005; SE = 0.0004; Z = -1.416; p 5 0.01). The observed size differences between poten- tially active and inactive cells were consistent with reports of stressed or starved cells being on average smaller than actively growing cells of the same species (reviewed in Kjelleberg et al. 1987).

Fig. 6. Potentially active (0) and inactive (v) cell sizes in Dio- patra cuprea tube lining biofllms. Mean (k SD) is shown for 1 Seasonal differences in cell parameters site in each tube. Sample size (n) for potentially active cells ranges from 43 to 94; sample size (n) for inactive cells ranges Comparisons were also made between mean cell from 33 to 79 numbers and sizes in Segment A of tubes collected in October and February 1997 There was no significant that infaunal burrows are sites of intense microbial difference in the mean number of cells per tube activity and contain substantial amounts of viable between the 2 sets (df 1,8; F = 0.31; p 2 0.50). How- microbial biomass (Aller & Yingst 1978, Aller 1980, ever, there was a difference in mean cell size. Mean Dobbs & Guckert 1988, Steward et al. 1996).What was cell sizes in tubes collected in October 1997 for activ- surprising was that cells at depth in the biofilm and at ity staining were significantly larger. 0.18 to 0.38 pm3, sites far below the sediment surface appeared to be as than those in tubes collected in February 1997, 0.058 capable of activlty as those near the biofilm surface to 0.09 pm3 (df 1,8;F = 0.093; p I 0.01). One clear dif- and at sites near the sediment surface. ference between the February and October samples Sizes of potentially active and inactive cells were was formaldehyde fixation of the February samples. also measured at 1 site in each of the 5 tube segments. Turley & Hughes (1992) reported about 10% reduc- The average size of potentially active cells ranged from tion in cell size (estimated from digitized cell images) 0.19 to 0.42 pm3 (grand mean = 0.3 pm3) and was sig- in fixed samples stored for 39 to 53 d, but provided no data for shorter storage times. Troussellier et al. (1995), using flow cytometric light scatter techniques, found no measurable changes in cell structure or mor- phology in samples fixed for a much shorter time period (24 h). The short times (I1 wk) our fixed sam- ples were stored prior to analysis seem unlikely to produce significant cell shrinkage. Even shrinkage comparab1.e to that reported by Turley & Hughes (1992) could not acco'unt for the almost 3-fold differ- ence in mean cell size we observed between the Feb- ruary and October samples. A more likely explanation for the difference in cell size can be based on the biol- ogy of Diopatra cuprea. The water temperature in October was warmer than in February and the poly- chaetes likely irrigated more actively in October (e.g. Mangum & Sassaman 1969). More frequent irrigation Cell size (pm3) during the warmer period would increase the avail- Fig. 7. Mean potentially active (m)and inactive (V) cell sizes in ability of oxygen and nutrients. Warmer water tem- Diopatra cuprea tube lining biofilms. Means (i SD) at each d.epth in the biofilm, across all 5 sites, is shown. The number perature would also have a positive effect on the bac- of samples at a given depth in the biofilm varies due to differ- terial communities, resulting In faster growth and an -ences in biofh thickness (n = 1 to 5) increase in mean cell size. Phillips & Lovell: Bacterial distributions in Diopatra cuprea tubes l??

The differences in the relationship of mean cell size Acknowledgements. We gratefully acknowledge John Kaig- and depth In the biofllm between the 2 collection ler for drawing Flg. 1, Holms Finch and Sarah Wood~n for dates could also be a result of irrigation intensity and their assistance with statistical analysis of data, Sarah Woodin for information on the lifestyle and habits of Diopatra cuprea, frequency. In February, became and and Courtney Richmond for her help with MorphoSys. We more uniform in size with increasing depth in the would also like to thank Bob Price (IMAF, Univ. South Car- biofilm. This trend was not observed in tubes col- olina Medical School) for his guidance In the use of the confo- lected in october and could be due to oxygen and cal microscope. This research was supported by grant number NO00149610403 from the United States Office of Naval nutrient limitation of microorganisms located deeper Research to C.R.L. in the biofilm.

LITERATURE CITED Additional observations Allen DM, Service SK, Ogburn-[Matthews MV (1992) Factors influencing the collection efficiency of estuarine fishes. Recent studies of many laboratory-cultivated bio- Trans Am Fish Soc 121:234-244 films have revealed channels of low-density-exo- Aller JY. Aller RC (1986) Evidence for localized enhancement polymer throughout the material (e.g. Stewart et al. of biological activity associated with tube and burrow structures in deep-sea sediments at the HEBBLE site, 1993, Stoodley et al. 1994, Wolfaardt et al. 1994). western North Atlantic. Deep-Sea Res 33:?55-790 These channels are thought to be essential for nutri- Aller RC (19801 Quantifyinq. - solute distributions in the bio- ent transport to organisms deep within the biofilm turbated zone of marine sediments by defining an average (Lawrence et al. 1994, Costerton et al. 1995) and can microenvironment. Geochim Cosmochim Acta 44: be revealed through staining with fluorescently 1955-1965 Aller RC (1983) The importance of the diffusive permeability tagged lectins, which bind specifically to the bio- of animal burrow linings in determining marine sediment film polysaccharides. Staining with fluorescein iso- chemistry. J Mar Res 41:299-322 thiocyanate-labeled concanavalin A (Sigma), which Aller RC (1988) Eenthic fauna and biogeochemical processes selectively binds a-mannopyranosyl and a-glucopy- in marine sedlments: the role of burrow structures. In: Blackburn TH, Sorensen J (eds) Nitrogen cycling in coastal ranosyl residues in polysaccharides, revealed no marine environments. Wilcy, Chichester, p 301-338 channels within the Diopatra cuprea tube biofilm Aller RC, Yingst JY (1978) Biogeochemistry of tube dwellings: (data not shown). D. cuprea deposits many layers of a study of the sedentary polychaete Amphitrite ornata mucus over time and cells that are on the surface of (Leidy). J Mar Res 36:201-254 the tube lining eventually become embedded within Binnerup SJ, Jensen K, Revsbech NP, Jensen MH, Sorensen J (1992) Denitrification, dissimilatory reduction of nitrate to the tube matrix. Any channels that might arise could ammonium, and nitrification in a bioturbated estuarine be filled with worm mucus on the next cycle of poly- sediment as measured with I5N and microsensor tech- mer deposition. However, channels may not be niques. Appl Environ Microbiol 58:303-313 required for active microbial metabolism throughout Boudreau BP, Marinelli RL (1994) A modelling study of dis- continuous biological irrigation J Mar Res 52:947-968 the tube-lining biofilm. Aller (1983) suggested that Chesbro W, Arbige M. Eifert R (1990) When nutrient limita- the tube linings of macro-infaunal organisms could tion places bacteria in the domalns of slow growth: meta- be considered to be highly hydrated porous mem- bolic, morphologic and cell cycle behavior. FEMS Microb branes (Aller 1983). Diffusion of small molecules Ecol?4:103-120 through the tube (including the biofilm) would not Costerton JW, Lewandowski 2, Caldwell DE, Korber DR. Lap- pin-Scott HM (1995) Microbial biofilrns. Annu Rev Micro- be hindered by such materials and thus activity biol 49:?11-745 of bacteria at depth in the biofilm could be sup- Dobbs FC, Guckert JB (1988) Callianassa trilobata (Crus- ported. tacea: Thalassinidae) influences abundance of melofauna The Diopatra cuprea tube supports an abundant and and biomass, composition, and physiological state of microbial communities within its burrow. Mar Ecol Prog varied microbial community that contains an unusually Ser 45:69-79 high proportion of potentially active cells. These find- Fenchel T (1969) The ecology of marine microbenthos IV. ings are consistent with the high levels of microbial Structure and function of the benthic ecosystem, its chem- biomass and activity reported for various infaunal bur- ical and physical factors and the microfauna communities rows and tubes. The trends observed in cell size distri- with special reference to the ciliated Protozoa. Ophelia 6: 1-182 butions are very intriguing in that they may reflect the Findlay RH, White DC (1983) The effects of feeding by the nutritional and/or stress regimes the organisms experi- sand dollar Mellita quinquiesperforata (Leske) on the ben- ence in response to irrigation of the tube by D. cuprea. thic microbial community. J Esp Mar Biol Ecol 72:25-41 Manipulative experiments exploring the impact of Findlay RH, Whlte DC (1984) In situ determination of metabolic activity in aquatic environments Microbiol Sci 1:90-95 tube irrigation on biofilm microbiota and molecular Findlay RH, Trexler MB, White DC (1990) Response of a ben- biological characterization of these organisms are thic microbial community to biotic disturbance. Mar Ecol ongoing. Prog Ser 62:135-148 Mar Ecol Prog Ser 183: 169-178, 1999

Grant J (1983) The relative magnitude of biological and phys- Norland S (1993) The relationship between biomass and vol- ical sediment reworking in an intertidal community. J Mar ume of bacteria. In: Kemp PF, Sherr BF, Sherr EB, Cole JJ Res 4 1:673-689 (eds) Handbook of methods in aquatic microbial ecology. Kaneshiro E, Wyder MA, Wu Y, Cushion MT (1993) Reliabil- Lewis Publishers, Boca Raton, FL, p 303-307 ity of calcein acetoxy methyl ester and etbdium homo- Novitsky JA (1983) Heterotrophic activity throughout a verb- dimer or propidium iodide for viability assessment of cal profile of seawater and sediment in Halifax Harbor, microbes. J Microbiol Methods 17:1- 16 Canada. Appl Environ Microbiol45:1753-1760 Kaprelyants AS, Gottschal JC, KellDB (1993) Dormancy in non- Novitsky JA, Karl DM (1986) Characterization of microbial sporulating bacteria. FEMS Mlcrobiol Rev 104:271-286 activity in the surface layers of a coastal sub-tropical sedi- Kjelleberg S, Hermansson M, Marden P, Jones GW (1987) ment. Mar Ecol Prog Ser 28:49-55 The transient phase between growth and nongrowth of Paerl HW, Pinckney JL (1996) A nuni-review of microbial con- heterotrophic bacteria, with emphasis on the marine envi- sortia: their roles in aquatic production and biogeochemi- ronment. Annu Rev Microbiol4 1:25-49 cal cycling. Microb Ecol 311225-247 Krantzberg G (1985) The influence of bioturbation on physi- Revsbech NP, Jorgensen BB (1986) Microelectrodes: their use cal, chemical and biological parameters in aquatic envi- in microbial ecology. Adv Microb Ecol9:293-352 ronments: a review. Environ Pollut 39:99-122 Roth BL, Poot M, Yue ST, Millard PJ (1997) Bacterial via- Kristensen E, Jensen MH, Andersen TK (1985) The impact of bility and antibiotic susceptibility testing with SYTOX polychaete (Nereis virens Sars) burrows on nitrification Green nucleic acid stain. Appl Environ Microbiol 63: and nitrate reduction in estuarine sediments. J Exp Mar 2421-2431 Biol Eco185:75-91 SAS Institute Inc (1997) SAS/STA~software: changes and Kr~stensenE, Jensen MH, Aller RC (1991) Direct measure- enhancements through release 6.12. SAS Institute Inc, ment of dssolved inorganic nitrogen exchange and deni- Cary, NC trification in individual polychaete (Nereis virens) bur- Steward CC, Nold SC, Ringelberg DB, White DC, Lovell CR rows. J Mar Res 493355-373 (1996) Microbial biomass and community structures in the Lawrence JR, Wolfaardt GM, Korber DR (1994) Determina- burrows of bromophenol producing and non-producing tion of dffusion coefficients in biofilms by confocal laser marine worms and surroundmg sediments. Mar Ecol Prog microscopy. Appl Environ 1Microbiol60:1166-1173 Ser 133:149-165 Lawrence JR, Korber DR, Wolfaardt GM, Caldwell DE (1997) Stewart PS, Peyton BM, Drury WJ, Murga R (1993) Quan- Analytical irnaging and microscopy techniques. In: Hurst titative observations of heterogeneities in Pseudomonas CJ, Knudsen GR, McInerney MJ. Stetzenbach LD, Walter aeruginosa biofilms. Appl Environ Microbiol 59:327-329 MV (eds) Manual of environmental microbiology. ASM Stoodley P, deBeer D, Lewandowski Z (1994) Liquid flow in Press, Washington, DC, p 29-51 biofilm systems. Appl Environ Microbiol60:2711-2716 Lebaron P, Parthuisot N, Catala P (1998) Comparison of blue Troussellier M, Courties C, Zettelmaier S (1995) Flow cyto- nucleic acid dyes for flow cytometric enumeration of metnc analysis of coastal lagoon bacterioplankton and bacteria in aquatic systems. Appl Environ Microbiol 64: picophytoplankton: fixation and storage effects. Estuar 1725-1730 Coast Shelf Sci 40:621-633 Lovell CR, Piceno Y (1994) Purification of DNA from estuarine Turley CM, Hughes DJ (1992) Effects of storage on direct esti- sediments. J Microbiol Methods 20:161-174 mates of bacterial numbers of preserved seawater sam- Mangum C, Sassaman C (1969) Temperature sensitivity of ples. Deep-Sea Res 39:375-394 active and inactive metabolism in a polychaetous annelid. Wolfaardt GM, Lawrence JR, Robarts RD. Caldwell SJ,Cald- Comp Biochem Physiol30:lll-116 well DE (1994) Multicellular organization in a degrada- Marinelli RL, Boudreau BP (1996) An experimental and mod- tive biofilm community. Appl Environ Microbiol 60: eling study of pH and related solutes in an irrigated anoxic 434-446 coastal sediment. J Mar Res 54:939-966 Yingst JY, Rhoads DC (1980) The role of bioturbation in the Meacham C, Duncan T (1991) MorphoSys. Version 1.29. Uni- enhancement of bacterial growth rates in marine sedi- versity Herbarium, Univerisity of California, Berkeley, CA ments. In: Tenore KR, Coull BC (eds) Marine benthic Myers AC (1972) Tube-worm-sediment relationships of Dio- dynamics. Univ South Carolina Press. Columbia, SC, patra cuprea (Polychaeta: Onuphidae) Mar Biol 17:350-356 p 407-421

Eoitolial responsibAty: Otto Kmne (Editor), Submitted September 4, 1998; Accepted: January 17, 1999 Oldendorf/Luhe, Germany Proofs received from author(s): June 4, 1999